Sumo Polaris II power amplifier Postscript: Turning the Other Cheek
After I sent the Sumo off to Santa Fe for measurements, I got a good news/bad news call from JA: "The Polaris you listened to was not only a pre-production sample, but defective as well (footnote 1). Its distortion on one channel was much higher than spec, and one of the TL-circuit trimpots had a bad wiper. A production sample is being Fed-Exed to you as we speak; could you give it a listen and write a followup by Monday?"
The good news? "I saw an Elvis movie on TV the other night!"
I also received a fax from Sumo's Mike Custer, who wanted to address some points I had made in my original review. For starters, the output wiring that spanned the full width of the board was a production error; future Polarises would have their output stages wired directly to the 5-way post closest to them. Also, the absence of the polystyrene bypass capacitor across the Mylar input-coupling cap was an error, too; production units would indeed include this bypass cap. Mike also stood up for his use of tin slip-on connectors in the Sumo's audio path and elsewhere, arguing that they were preferable to good soldered connections. I disagree with this; I've found that a correctly soldered connection, using Wonder Solder or the equivalent, is superior to many gold-gold connections, much less a junction of tin-tin as found at numerous points in the Polaris.
Well, Mike was correct on the two construction revisions: removing the cover revealed that both the output wiring correction and the polystyrene bypass cap (0.01µF) were there. However, as the output wiring effectively swapped the left/right identities of the output connectors, the red and white plastic inserts identifying the input RCA jacks were now wrong! Sumo has silkscreened the letters A and B by both the input and output connectors, and a signal applied to input A will result in an output signal at binding post A, but as the left RCA jack is red and the right white, I suspect that many users would hook up their system with the channels reversed. I'm sure Sumo will correct this, but for now, it's very confusing.
There were two construction problems with the second Sumo. First, there was a great deal more power-transformer buzz emanating from the chassis than the first amp; removing the cover revealed this to be due to a transformer mounting screw missing from one corner, which let the transformer vibrate more freely against the aluminum chassis. Second, a large blob of brown translucent glue engulfed the open back of the pc-mounted input RCA connector assembly, ostensibly to brace it against the board. This glue bridged the four adjacent conductors (signal left and right, ground left and right) along their entire length from jack to board, acting as an additional dielectric. This is most undesirable, and an anomaly that was absent from the first, pre-production sample.
As the original Polaris wasn't available for direct comparison as was the case with the two Counterpoints I received—JA had broken it while doing his measurements—I used the Adcom GFA-555 II as a sonic yardstick by which to judge the new Sumo. It was immediately obvious that this Polaris bore very little sonic resemblance to the pre-production sample. While the grain in the highs was much lower (although still greater than the Adcom), the midrange sounded quite smeared and veiled. While the midrange of the first Sumo had a very forward character but still allowed spatial information to be decoded well, this second Polaris blurred the soundstage to a significant degree. The Stanczyk piano recording on the Stereophile Test CD reproduced more like an amorphous assortment of notes, rather than as a distinctly outlined piano. With the Adcom, it was easy to pick out each individual key on the piano; the Sumo blurred these locations sufficiently to make the piano's image difficult, if not outright impossible, to make out. On the plus side, the midrange exhibited little of the lower-midrange accentuation and upper-midrange attenuation of the first sample; however, given the choice, I'd take the tonal-balance colorations of the latter over the image inspecificity of the former any day.
And now for something completely different: a Sumo with no buttcheeks! The bass of this Polaris was altogether changed from the first amp; where the original Polaris competed quite favorably with the outstanding bass of the Adcom, Sumo II just wasn't up to the fight. The low end was neither as tight nor as extended as the Adcom, lending JA's piano recording a much smaller sound. Bass-happy CDs like Lenny Kravitz's Let Love Rule (Virgin 91290-2) sounded less powerful than on the Adcom, with greatly reduced impact and extension.
In short, this second amp has me thoroughly puzzled; while one area has been greatly improved—ie, the reduction of high-end grain—the others have gotten worse. And in the case of the midrange, a lot worse. I have no idea why the malfunctioning amp had better bass and midrange performance, and significantly better soundstaging, but that's what I heard.
The first, defective Polaris was almost a legitimate alternative to the Adcom GFA-555 II, but this production sample falls well short of the mark. While I stated in my review that prospective buyers in this price range might include the first Sumo Polaris in their auditions if a more forward tonal balance is desirable, I can't recommend the second sample I received. Its questionable sonics, coupled with the construction defects of the missing power-transformer mounting screw and the glue blob in the RCA block, remove it from my consideration.—Corey Greenberg
Footnote 1: JA has assured me that this situation is a sad fact of life with reviewing, but I think sending a pre-production version—ie, a unit that may not be identical or even similar to what you, the reader, can buy in the store—is terribly unfair, to both of us. It's unfair to me because I end up spending a great deal of time and effort listening to and writing about a unit that may have nothing to do with the actual version on dealers' shelves, and it's unfair to you because my review is ultimately of little worth to you as an accurate account of the product. It's one thing for a manufacturer to accidentally send out a defective unit for review; submitting a prototype or pre-production sample is wrong, plain and simple. Of course, were I to give an unqualified rave to a pre-production sample, surely the manufacturer would still take it upon himself to dash off an actual production sample for a followup, right?—Corey Greenberg